US20110170553A1 - Method of data delivery across a network fabric in a router or ethernet bridge - Google Patents

Method of data delivery across a network fabric in a router or ethernet bridge Download PDF

Info

Publication number
US20110170553A1
US20110170553A1 US12990370 US99037009A US20110170553A1 US 20110170553 A1 US20110170553 A1 US 20110170553A1 US 12990370 US12990370 US 12990370 US 99037009 A US99037009 A US 99037009A US 20110170553 A1 US20110170553 A1 US 20110170553A1
Authority
US
Grant status
Application
Patent type
Prior art keywords
ethernet
data
network
protocol
fabric
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
US12990370
Other versions
US9401876B2 (en )
Inventor
Jon Beecroft
David Charles Hewson
Anthony Michael Ford
Mark Owen Homewood
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Cray UK Ltd
Original Assignee
Gnodal Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L12/00Data switching networks
    • H04L12/28Data switching networks characterised by path configuration, e.g. local area networks [LAN], wide area networks [WAN]
    • H04L12/46Interconnection of networks
    • H04L12/4633Interconnection of networks using encapsulation techniques, e.g. tunneling
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/30Peripheral units, e.g. input or output ports
    • H04L49/3009Header conversion, routing tables or routing tags
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L49/00Packet switching elements
    • H04L49/35Application specific switches
    • H04L49/351LAN switches, e.g. ethernet switches
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L2212/00Encapsulation of packets

Abstract

The invention provides an Ethernet bridge or router comprising a network fabric adapted to provide interconnectivity to a plurality of Ethernet ports, each of the Ethernet ports being adapted to receive and/or transmit Ethernet frames, and wherein the Ethernet bridge or router further comprises an encapsulator connected to receive Ethernet Protocol Data Units from the Ethernet ports, wherein the encapsulator is operable to generate a Fabric Protocol Data Unit from a received Ethernet Protocol Data Unit, the Fabric Protocol Data Unit comprising a header portion, and a payload portion which comprises the Ethernet Protocol Data Unit concerned, and wherein the encapsulator is operable to transform Ethernet destination address information from the Ethernet Protocol Data Unit into a routing definition for the network fabric, and to include this routing definition in the header portion of the Fabric Protocol Data Unit. Also provided is a method of data delivery across a network.

Description

    BACKGROUND Technical Field of the Invention
  • [0001]
    The present invention generally relates to an Ethernet bridge or router and to a method of data delivery across a network suitable for use in, but not limited to, multi-processor networks such as storage networks, data centres and high performance computing. In particular, the present invention is concerned with an Ethernet bridge or router which is adapted for the distribution of standard IEEE 802 data frames or data frames meeting future Ethernet standards.
  • Protocol Layers
  • [0002]
    Conceptually, an Ethernet network is decomposed into a number of virtual layers in order to separate functionality. The most common and formally standardised model used is the Open Systems Interconnect (OSI) reference model. A useful article which described in detail the OSI reference model is “OSI Reference Model—The ISO Model of Architecture for Open Systems Interconnection” by Hubert Zimmermann, IEEE Transactions on Communications, Vol. COM-28, No. 4, April 1980. The OSI reference model comprises seven layers of network system functionality, as follows:
      • 1. Physical Layer is responsible for physical channel access. It consists of those elements involved in transmission and reception of signals, typically line drivers and receivers, signal encoders/decoders and clocks.
      • 2. Data Link Layer provides services allowing direct communication between end-station devices over the underlying physical medium. This layer provides Framing, separating the device messages into discrete transmissions or frames for the physical layer, encapsulating the higher layer packet protocols. It provides Addressing to identify source and destination devices. It provides Error Detection to ensure that corrupted data is not propagated to higher layers.
      • 3. Network Layer is responsible for network-wide communication, routing packets over the network between end-stations. It must accommodate multiple Data Link technologies and topologies using a variety of protocols, the most common being the Internet Protocol (IP).
      • 4. Transport Layer is responsible for end-to-end communication, shielding the upper layers from issues caused during transmission, such as dropped data, errors and mis-ordering caused by the underlying medium. This layer provides the application with an error-free, sequenced, guaranteed delivery message service, managing the process to process data delivery between end stations. Transmission Control Protocol (TCP) and User Datagram Protocol (UDP) are the most commonly recognised Transport Layer protocols.
      • 5. Session Layer is responsible for establishing communications sessions between applications, dealing with authentication and access control.
      • 6. Presentation Layer ensures that different data representations used by machines are resolved.
      • 7. Application Layer provides generic functions that allow user applications to communicate over the network.
  • [0010]
    For the purposes of this discussion we need not consider operations above the Transport Layer as it should, if well implemented, shield higher layers from issues arising in and below its scope.
  • Data Unit Encapsulation
  • [0011]
    As data passes from a user application down through the protocol layers, each layer adds some control information to the data transmission so that the application's peer can properly interpret it at its destination. This process is called ‘encapsulation’, with each layer taking the information and structure provided to it by a higher layer and adding further layer specific control, as defined by the protocol a layer implements, typically, protocol header and trailers are calculated and added. This combined, layer specific Protocol Data Unit (PDU) is thus passed down the layer stack. Similarly, once the data is received at the destination, layer specific protocol information is stripped from the PDU and interpreted before the PDU is passed on to higher layers in the stack, ultimately arriving in the peer application.
  • [0012]
    PDUs have specific names at each layer in the stack:
      • 1. Physical Layer refers to the PDU as a symbol stream.
      • 2. Data Link Layer refers to the PDU as a frame.
      • 3. Network Layer refers to the PDU as a packet or datagram.
      • 4. Transport Layer refers to the PDU as a segment or message
    Network Interconnections
  • [0017]
    A device that implements network services at the Data Link Layer and above is called a ‘station’. The Physical Layer is excluded from this definition as it is not addressable by a protocol. There are two types of station:
      • End Stations are the ultimate source and destination of network data communicated across a network.
      • Intermediate Stations forward network data generated by End Stations between source and destination. An intermediate station which forwards completely at the Data Link Layer is commonly called a ‘Bridge’; a station which forwards at the Network Layer is commonly called a ‘Router’.
    Ethernet Routing
  • [0020]
    An intermediate station performs routing operations directing PDUs that appear on one ingress port to another egress port. An intermediate station can operate either at the Data Link Layer, in the case of a Layer 2 intermediate station, or at the Network Layer, in the case of a Layer 3 intermediate station. Layer 2 intermediate stations use the Media Access Control (MAC) address, embedded in the Data Link Layer PDU encapsulation, to control the routing operation. Layer 3 intermediate stations use the IP address, embedded in the Network Layer PDU encapsulation, to direct routing. Layer 3 routing performs a number of steps resulting in the modification of both the source and destination Ethernet segment MAC addresses and thus allows the egress port of the intermediate station to be determined. There are additional operations that need to be performed at ingress in addition to routing to enable Ethernet features such as port security, reserved MAC addresses, VLAN or link aggregate membership. This lengthy sequence of operations is performed at every intermediate station. These operations are costly, usually requiring a large translation mechanism with a number of complex rules to govern the control and management of the translation entries held in the translation tables. The translation mechanism must be capable of sustaining a very high throughput of translations in order to keep up with many ports operating at very high throughput rates. The checking and translation operations can take a great amount of time significantly adding to delivery latency.
  • [0021]
    As already stated a large network may be constructed from many intermediate stations (bridges or routers) and PDUs being delivered from one end station to another may have to pass through many intermediate stations to complete their delivery. Every intermediate station the PDU passes through must complete the complex address checking and translation operation. This can significantly slow delivery of the frame and increase the power consumption of the whole system.
  • DESCRIPTION OF RELATED ART
  • [0022]
    FIG. 1 illustrates a conventional Ethernet bridge 1 with its ports 3 connected to a plurality of individual Ethernet stations 2. In overview, the Ethernet bridge 1 comprises a plurality of ports 3, means for performing Ethernet receive functions 4, means for performing Ethernet transmit functions 5, a packet buffer 6 a crossbar 7. In FIG. 1 only one port 3 is illustrated. However, in practice, each connection to a separate Ethernet station has a respective port with respective receive and transmit functionality.
  • [0023]
    In US 2004/0131073 a modular scalable switch adapted for the distribution of fast Ethernet frames is described. The Ethernet switch comprises a plurality of individually programmable one-port communication modules which access a common distribution bus. Although described as scalable, this and other Ethernet switches consisting of many separate low port count modules that are dependent upon a common bus, are fundamentally unsuited to networks involving very large numbers of processors such as storage centres and server farms. The number of ports which can be supported within an individual bridge is limited by the bus capacity of these architectures. As the density, capacity and data rates of individual connections increases these common bus architectures limit the construction of large Ethernet networks, bus bandwidth quickly being exceeded by a relatively small number of high performance connections. The present invention seeks to overcome the disadvantages with the above Ethernet switch and other known Ethernet switches and seeks to provide an Ethernet bridge or router and a method of data delivery across a network fabric suitable for use in, but not limited to, multi-processor networks such as storage networks, data centres and high performance computing.
  • SUMMARY OF THE INVENTION
  • [0024]
    The present invention therefore provides a single Ethernet bridge or router comprising a network fabric adapted to provide interconnectivity to a plurality of Ethernet ports, each of the Ethernet ports being adapted to receive and/or transmit Ethernet frames, and wherein the Ethernet bridge or router further comprises an encapsulator connected to receive Ethernet Protocol Data Units from the Ethernet ports, wherein the encapsulator is operable to generate a Fabric Protocol Data Unit from a received Ethernet Protocol Data Unit, the Fabric Protocol Data Unit comprising a header portion, and a payload portion which comprises the Ethernet Protocol Data Unit concerned, and wherein the encapsulator is operable to transform Ethernet destination address information from the Ethernet Protocol Data Unit into a routing definition for the network fabric, and to include this routing definition in the header portion of the Fabric Protocol Data Unit.
  • [0025]
    In a further aspect the present invention provides a method of data delivery across a network comprising a network fabric adapted to provide interconnectivity to a plurality of Ethernet ports, each of the ports being adapted to receive and/or transmit Ethernet data frames, the method comprising the steps of:
      • receiving an Ethernet frame or packet;
      • generating a Fabric Protocol Data Unit from the received Ethernet Protocol Data Unit, the Fabric Protocol Data Unit comprising a header portion and a payload portion, wherein the payload portion comprises Ethernet destination address information and the header portion comprises a routing definition for the network fabric derived from the Ethernet destination address;
      • transmitting the Fabric Protocol Data Unit from an ingress network port of the network fabric to at least one egress network port of the network fabric;
      • extracting the Ethernet frame or packet from the Fabric Protocol Data Unit; and
      • delivering the Ethernet frame or packet to an Ethernet device.
  • [0031]
    In a yet further aspect the present invention provides an Ethernet bridge or router comprising a network fabric adapted to provide interconnectivity to a plurality of Ethernet ports, each of the Ethernet ports being adapted to receive and/or transmit Ethernet frames, and wherein the Ethernet bridge or router further comprises software instructions for operating an encapsulator to generate a Fabric Protocol Data Unit from a received Ethernet Protocol Data Unit, the Fabric Protocol Data Unit comprising a header portion, and a payload portion which comprises the Ethernet Protocol Data Unit concerned, and wherein the encapsulator is operable to transform Ethernet destination address information from the Ethernet Protocol Data Unit into a routing definition for the network fabric, and to include this routing definition in the header portion of the Fabric Protocol Data Unit.
  • [0032]
    In a preferred embodiment the network fabric comprises a plurality of switches and a plurality of network ports adapted to receive and/or transmit the FPDU directly to an egress network port, without de-encapsulation or re-encapsulation of the Ethernet PDU whereby the Ethernet PDU of the FPDU is invisible to the switches of the network bridge fabric.
  • [0033]
    Ideally, the network fabric is adapted so as not to modify any part of the payload of the FPDU.
  • [0034]
    Preferably, each one of the network ports is allocated a destination number which is representative of the physical position of the network port on the network fabric whereby algorithmic routing of the FPDU across the network is enabled.
  • [0035]
    More preferably, each of the plurality of Ethernet ports is programmable and adapted to support automatically both Ethernet PDUs and proprietary FPDUs, the ports automatically interpreting either protocol when it is received.
  • [0036]
    With the present invention, the encapsulation means implements a new protocol layer additional to the protocol layers of the OSI model for Ethernet networks, the new protocol layer being stacked between the Physical Layer and the Data Link Layer of the OSI model for Ethernet networks and provides for encapsulation of network layer PDUs and data link layer PDUs in the FPDU.
  • [0037]
    In a further preferred embodiment the encapsulation means is adapted to allow for the payload of an FPDU to be interrupted for the insertion of one or more control tokens.
  • [0038]
    In a further preferred embodiment the encapsulation means is adapted to allow for the removal or replacement of control tokens previously inserted into the payload of an FPDU.
  • [0039]
    In the context of this document reference to an intermediate station is intended to also encompass a station capable of forwarding a protocol encapsulation at a point between the Data Link Layer and the Physical Layer. This type of intermediate station will be referred to as a ‘Bridge Fabric’ later in this document. A multi-port Bridge Fabric may be implemented by a collection of ‘Bridge Fabric Switches’ (BFS) interconnected by ‘Bridge Fabric Switch Links’ (BFSL).
  • [0040]
    Thus, with the present invention a method for Ethernet encapsulation is provided which enables improved routing and data delivery across a network fabric, without change to the underlying Ethernet protocol structures or contents. This provides, amongst other things, the mechanism to guarantee the order of Ethernet data delivered across such a network that includes multiple paths between source and destination endpoints, completely removing the possibility of mis-ordering or duplication being propagated by the network, enabling the construction of very high port count Ethernet Bridges and Routers.
  • [0041]
    The foregoing and other objects, aspects and advantages will be better understood from the following detailed description of a preferred embodiment of the present invention with reference to the drawings, in which:
  • BRIEF DESCRIPTION OF THE DRAWINGS
  • [0042]
    FIG. 1 illustrates schematically an Ethernet bridge of the prior art;
  • [0043]
    FIG. 2 illustrates schematically an Ethernet bridge in accordance with the present invention;
  • [0044]
    FIG. 3 illustrates schematically a network fabric element used in the Ethernet bridge of FIG. 2;
  • [0045]
    FIG. 4 is a flow diagram of a method of data delivery across a network in accordance with the present invention;
  • [0046]
    FIG. 5 illustrates the format of an FPDU used in the method in accordance with the present invention; and
  • [0047]
    FIGS. 6 and 7 illustrate the format of control tokens used in the method in accordance with the present invention.
  • DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT
  • [0048]
    The Ethernet bridge or router described below introduces an additional protocol layer, referred to herein as an ‘Encapsulation Layer’, that appears between the Physical Layer and the Data Link Layer of the standard OSI model which can encapsulate both Network Layer and Data Link Layer PDUs.
  • [0049]
    Complete encapsulation in a PDU unique to the present invention, which is referred to herein as a ‘Fabric Protocol Data Unit’ (FPDU), avoids the necessity of modifying the underlying PDU frame headers or trailers, and thus removes the overhead of recalculating the cyclic redundancy check (CRC) or other derived information based upon the contents of the frame. A FPDU is used in implementation of the data transmission, acknowledgement and flow-control mechanisms and can be further utilised to provide many other attractive features important to large high performance, scalable Ethernet networks.
  • [0050]
    An Ethernet bridge or router 10 is illustrated in FIG. 2 which may be connected to a plurality of separate Ethernet stations 2 and which implements the encapsulation of both Network Layer and Data Link Layer PDUs in a FPDU. The Ethernet bridge 10 generally comprises a network 12 in combination with a plurality of Ethernet ports 11 (only one is illustrated for the sake of clarity) with each port being individually connectable to an Ethernet station. The Ethernet ports 11 are generally conventional in design and each includes means for establishing a data connection with an Ethernet station, a receiving means 13 for performing Ethernet receive functions 13 and a transmitting device 14 for performing Ethernet transmit functions.
  • [0051]
    The Ethernet ports 11 are connected to a network interface 15 which provides conventional functionality such as packet buffering 16. However, the network interface 15 additionally includes an Ethernet PDU encapsulator 17 which connects the network interface 15 to ingress ports (not shown) of the network 12 and an Ethernet PDU decapsulator 18 which connects egress ports (not shown) of the network 12 back to the Ethernet ports 11. The Ethernet PDU encapsulator 17 implements the protocol of the Encapsulation Layer and thus is responsible for the generation of the FPDUs. Ideally, each port 11 of the Ethernet bridge 10 has a respective network interface 15 and thus a respective Ethernet PDU encapsulator 17 and a respective Ethernet PDU decapsulator 18.
  • [0052]
    In the embodiment of the Ethernet bridge 10 illustrated in FIG. 2 all of the ports of the Ethernet bridge 10 are Ethernet ports. However, it is envisaged that the ports of the Ethernet bridge 10 may be programmable to function as either Ethernet ports or proprietary ports which provide for connection of the bridge 10 to non-Ethernet compatible devices.
  • [0053]
    The network 12, which is a proprietary network, comprises a plurality of interconnected network fabric elements 19 in the form of bridge fabric switches (BFSs) interconnected by bridge fabric switch links (BFSLs). The details of the interconnections of the BFSs are not of relevance to the present invention. In FIG. 3 the structure of one of the network fabric elements 19 is illustrated in greater detail. Thus the network fabric element 19 includes a crossbar 22 the input to which is fed via means for performing frame buffering 20 and means for selecting a crossbar route. The network fabric element 19 may, of course, include additional processing elements as required by the proprietary network 12.
  • [0054]
    A method of data delivery across the network fabric 12 which implements the protocol of the Encapsulation Layer is illustrated in FIG. 4. Encapsulation occurs at ingress to the Ethernet bridge 10. An Ethernet PDU transmitted by an Ethernet station 2 is received S1 at a port 11 of the Ethernet bridge 10 and is communicated via the receiving device 13, which performs conventional Ethernet receiving functions, to the network interface 15 and in particular the Ethernet PDU encapsulator 17. The encapsulator 17 attaches S2 a FPDU header to the Data Link Layer frame. Also, the encapsulator 17 extracts addressing information from within the Ethernet PDU and transforms the addressing information into a destination descriptor S3 that identifies the complete route through the network fabric 12 to the required destination for the PDU. This destination descriptor is then added to the FPDU header. The data is then communicated S4 to an ingress port of the network fabric 12 and is transmitted across the network 12 via a plurality of network fabric elements 19 to an egress port of the network. At egress from the network 12 the FPDU information is stripped S5 by the Ethernet PDU decapsulator 19 so that only the original Ethernet PDU remains. The unmodified Ethernet PDU is then forwarded S6 to an Ethernet station which is the required destination for the PDU. Thus, the protocol and mechanisms enabling transport of the data across the network 12 are completely contained within the network and are invisible to the Ethernet stations at both ingress and egress.
  • [0055]
    FPDU encapsulation of the entire Ethernet PDU enables significantly improved speed of delivery of Ethernet frames across the network fabric, and reduced latency, because the Ethernet PDU content is not required to be inspected or modified in any way during delivery across the network. This comes from simplified routing and the possibility to introduce additional delivery controls to maximise the utilisation of the bridge fabric infrastructure.
  • [0056]
    Part of the new encapsulation may include control tokens. These tokens are multiplexed onto the BFSL along with the encapsulated Ethernet PDU. That is to say, the tokens may interrupt the payload of an FPDU which is in the process of being delivered across the network. Thus encapsulation of the Ethernet PDU also improves the delivery latency of the control tokens by allowing them to be inserted within the encapsulated Ethernet PDU. A large PDU can take microseconds to be transmitted. Allowing the tokens to briefly interrupt the encapsulated Ethernet PDU has little effect on the delivery time of the FPDU but reduces the delivery time of the token from multiple microseconds to tens of nanoseconds.
  • [0057]
    The control tokens are used to give additional information that help with the delivery of the encapsulated Ethernet PDUs. This includes line flow control and Ethernet delivery acknowledgement. It can also be used for sideband operations such as fetching address translations, verifying the fabric has been correctly cabled up and exchanging congestion information. The content of an FPDU is illustrated in FIG. 5. In the case of FIG. 5 the FPDU comprises a payload 30 consisting of an interruptible Ethernet packet. The header 31 to the payload 30 comprises various network specific data including, but not limited to, a start token 32, a destination address 33, the size of the payload 34 and CRC 35. Having a CRC in the FPDU header guarantees the header's validity allowing it to be used without having to wait for the rest of the encapsulated payload to arrive. Various different types of control tokens may, of course, be implemented for use with the network described herein. Furthermore, such control tokens may range from one or two word tokens up to five or more word tokens depending upon the control functionality required. In FIGS. 6 and 7 two different types of control token which are suitable for use with the Ethernet bridge described herein, are illustrated.
  • [0058]
    A PDU will have to cross many links in a large BFS. Each time the PDU arrives at a new bridge or router the new destination port must be determined quickly from the route information. As mentioned earlier, for Ethernet protocols the route information is held in either the MAC address at Layer 2 or the IP address at Layer 3 and as already stated the translation process is complex and can significantly add to the routing latency. Adding the new encapsulation allows the translation of the MAC address or the IP address to be performed only once for each PDU crossing the BFS.
  • [0059]
    Usually there is little or no correlation between the physical location of end stations and the addresses used to identify them within the network. This is true when the end station is identified by an IP or MAC address. The new encapsulation includes a destination descriptor or route identifier, the value of which has been assigned to correlate to the physical locations of the Ethernet ports of the entire BFS. This enables interval routing to be performed and means that a PDU crossing the network can use a simple algorithmic approach to selecting an output port. The complex translation of the MAC and IP address only needs to be performed once on ingress to the BFS.
  • [0060]
    The Ethernet bridge or router described herein is truly scalable offering from 256 ports or fewer up to 48,000 ports or more. A single Ethernet bridge or router using the method described herein is capable of providing greatly increased connectivity in comparison to conventional Ethernet bridges. For example, currently the largest 10 Gbe Ethernet bridges (which are modular in construction) offer only 288 ports. With the Ethernet bridge of the present invention, which is capable of operating at 10 Gbe or above, a single bridge is capable of offering 48,000 ports.
  • [0061]
    It is to be understood that various modifications to the preferred embodiment and the generic principles and features described herein will be readily apparent to those skilled in the art. Thus, the present invention is not intended to be limited to the embodiment shown and such modifications and variations also fall within the spirit and scope of the appended claims.

Claims (21)

  1. 1. An Ethernet bridge or router comprising a network fabric adapted to provide interconnectivity to a plurality of Ethernet ports, each of the Ethernet ports being adapted to receive and/or transmit Ethernet frames, the Ethernet bridge or router further comprising an encapsulator connected to receive Ethernet Protocol Data Units from the Ethernet ports, the encapsulator being operable to generate a Fabric Protocol Data Unit from a received Ethernet Protocol Data Unit, the Fabric Protocol Data Unit comprising a header portion, and a payload portion which comprises the Ethernet Protocol Data Unit concerned, wherein the encapsulator is operable to transform Ethernet destination address information from the Ethernet Protocol Data Unit into a routing definition which defines a complete route for the Fabric Protocol Data Unit through the network fabric, and is operable to include this routing definition in the header portion of the Fabric Protocol Data Unit.
  2. 2. The Ethernet bridge or router of claim 1, wherein the network fabric comprises a plurality of switches and a plurality of network ports adapted to receive and/or transmit the Fabric Protocol Data Unit directly to an egress network port, without de-encapsulation or re-encapsulation of the Ethernet Protocol Data Unit.
  3. 3. The Ethernet bridge or router of claim 2, wherein the Ethernet Protocol Data Unit of the Fabric Protocol Data Unit is invisible to the switches of the network fabric.
  4. 4. The Ethernet bridge or router of claim 1, wherein the network fabric is adapted so as not to modify any part of the payload of the Fabric Protocol Data Unit.
  5. 5. The Ethernet bridge or router of claim 2, wherein each of the network ports is allocated a destination number which is representative of the physical position of the network port on the network fabric thereby enabling algorithmic routing of the Fabric Protocol Data Unit across the network.
  6. 6. The Ethernet bridge or router of claim 1, wherein the plurality of Ethernet ports are programmable and adapted to support automatically both ethernet Protocol Data Units and proprietary Fabric Protocol Data Units, the ports automatically interpreting either protocol when it is received.
  7. 7. The Ethernet bridge or router claim 1, wherein the encapsulator implements a new protocol layer additional to the protocol layers of the Open Systems Interconnect model for Ethernet networks.
  8. 8. The Ethernet bridge or router as claimed in claim 7, wherein the new protocol layer is stacked between the Physical Layer and the Data Link Layer of the Open Systems Interconnect model for Ethernet networks and provides for encapsulation of network layer Protocol Data Units and data link layer Protocol Data Units in the Fabric Protocol Data Unit.
  9. 9. The Ethernet bridge or router of claim 1, wherein the encapsulator is adapted to allow for the payload of a Fabric Protocol Data Unit to be interrupted for the insertion of one or more control tokens
  10. 10. A method of data delivery across a network comprising a network fabric adapted to provide interconnectivity to a plurality of Ethernet ports, the method comprising the steps of:
    receiving an Ethernet frame or packet;
    generating a Fabric Protocol Data Unit from the received Ethernet Protocol Data Unit, the Fabric Protocol Data Unit comprising a header portion and a payload portion, wherein the payload portion comprises Ethernet destination address information and the header portion comprises a routing definition for the network fabric derived from the Ethernet destination address;
    transmitting the Fabric Protocol Data Unit from an ingress network port of the network fabric to at least one egress network port of the network fabric;
    extracting the Ethernet frame or packet from the Fabric Protocol Data Unit; and
    delivering the Ethernet frame or packet to an Ethernet device.
  11. 11. The method of claim 10, wherein the network fabric comprises a plurality of switches and a plurality of network ports wherein during the step of transmitting the Fabric Protocol Data Unit the Ethernet Protocol Data Unit is neither de-encapsulated or re-encapsulated by any of the switches or network ports.
  12. 12. The method of claim 11, wherein during the step of transmitting the Fabric Protocol Data Unit the Ethernet Protocol Data Unit of the Fabric Protocol Data Unit is invisible to the switches of the network fabric.
  13. 13. The method of claim 12, wherein during the step of transmitting the Fabric Protocol Data Unit no part of the payload of the Fabric Protocol Data Unit is modified.
  14. 14. The method of claim 11, wherein each one of the network ports is allocated a destination number which is representative of the physical position of the network port on the network fabric whereby the Fabric Protocol Data Unit is transmitted across the network using algorithmic routing.
  15. 15. The method of claim 10, wherein the plurality of Ethernet ports are programmable and adapted to support automatically both Ethernet Protocol Data Units and proprietary Fabric Protocol Data Units whereby during the step of receiving an Ethernet frame the Ethernet ports automatically interpret either protocol.
  16. 16. The method of claim 10, wherein the step of generating a Fabric Protocol Data Unit implements a new protocol layer additional to the protocol layers of the Open Systems Interconnect model for Ethernet networks.
  17. 17. The method of claim 16, wherein the new protocol layer is stacked between the Physical Layer and the Data Link Layer of the Open Systems Interconnect model for Ethernet networks and provides for encapsulation of network layer Protocol Data Units and data link layer Protocol Data Units in the Fabric Protocol Data Unit.
  18. 18. The method of claim 10, further comprising the step of interrupting the generation of a Fabric Protocol Data Unit, adding one or more control tokens to the Fabric Protocol Data Unit and transmitting the Fabric Protocol Data Unit with the one or more control tokens across the network fabric.
  19. 19. The method of claim 18, further comprising the step of replacing or removing one or more control tokens previously inserted into a Fabric Protocol Data Unit.
  20. 20. An Ethernet bridge or router comprising a network fabric adapted to provide interconnectivity to a plurality of Ethernet ports, each of the Ethernet ports being adapted to receive and/or transmit Ethernet frames, and wherein the Ethernet bridge or router further comprises software instructions for operating an encapsulator to generate a Fabric Protocol Data Unit from a received Ethernet Protocol Data Unit, the Fabric Protocol Data Unit comprising a header portion, and a payload portion which comprises the Ethernet Protocol Data Unit concerned, and wherein the encapsulator is operable to transform Ethernet destination address information from the Ethernet Protocol Data Unit into a routing definition for the network fabric, and to include this routing definition in the header portion of the Fabric Protocol Data Unit.
  21. 21. The Ethernet bridge or router of claim 1, wherein the encapsulator is operable to transform Ethernet destination address information from the Ethernet Protocol Data Unit into a routing definition which defines a set of complete routes for the Fabric Protocol Data Unit through the network fabric.
US12990370 2008-05-01 2009-04-29 Method of data delivery across a network fabric in a router or Ethernet bridge Active 2029-11-30 US9401876B2 (en)

Priority Applications (3)

Application Number Priority Date Filing Date Title
GB0807937A GB2459838B (en) 2008-05-01 2008-05-01 An ethernet bridge and a method of data delivery across a network
GB0807937.8 2008-05-01
PCT/GB2009/001091 WO2009133369A8 (en) 2008-05-01 2009-04-29 A method of data delivery across a network fabric in a router or ethernet bridge

Publications (2)

Publication Number Publication Date
US20110170553A1 true true US20110170553A1 (en) 2011-07-14
US9401876B2 US9401876B2 (en) 2016-07-26

Family

ID=39537117

Family Applications (2)

Application Number Title Priority Date Filing Date
US12990370 Active 2029-11-30 US9401876B2 (en) 2008-05-01 2009-04-29 Method of data delivery across a network fabric in a router or Ethernet bridge
US15218762 Pending US20160337146A1 (en) 2008-05-01 2016-07-25 Method of data delivery across a network fabric in a router or ethernet bridge

Family Applications After (1)

Application Number Title Priority Date Filing Date
US15218762 Pending US20160337146A1 (en) 2008-05-01 2016-07-25 Method of data delivery across a network fabric in a router or ethernet bridge

Country Status (5)

Country Link
US (2) US9401876B2 (en)
EP (1) EP2291959B1 (en)
CN (1) CN102318283A (en)
GB (1) GB2459838B (en)
WO (1) WO2009133369A8 (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3185491A1 (en) * 2014-02-05 2017-06-28 INTEL Corporation Reliable transport of ethernet packet data with wire-speed and packet data rate match

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB2459838B (en) * 2008-05-01 2010-10-06 Gnodal Ltd An ethernet bridge and a method of data delivery across a network
GB2461132B (en) * 2008-06-27 2013-02-13 Gnodal Ltd Method of data delivery across a network
CN104767590A (en) * 2015-04-08 2015-07-08 江苏飞尚安全监测咨询有限公司 Reliable data transmission and control method for serial communication

Citations (30)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5872783A (en) * 1996-07-24 1999-02-16 Cisco Systems, Inc. Arrangement for rendering forwarding decisions for packets transferred among network switches
US5999531A (en) * 1998-04-17 1999-12-07 Cabletron Systems, Inc. Method and system for identifying ports and forwarding packets in a multiport switch
US20010043609A1 (en) * 1998-04-28 2001-11-22 Alan Stanley John Chapman Efficient packet data transport mechanism and an interface therefor
US20020018475A1 (en) * 1998-07-22 2002-02-14 Yoram Ofek Switching with multiple time references
US20020061018A1 (en) * 2000-10-03 2002-05-23 Altima Communications, Inc. Switch transferring data using data encapsulation and decapsulation
US20020083190A1 (en) * 2000-12-26 2002-06-27 Satoshi Kamiya Apparatus and method for GFP frame transfer
US20020101842A1 (en) * 2001-02-01 2002-08-01 Harrison Lee C. Interface system for wireless node and network node
US20020165978A1 (en) * 2001-05-07 2002-11-07 Terence Chui Multi-service optical infiniband router
US20030152182A1 (en) * 2000-08-22 2003-08-14 Pai B. Anand Optical exchange method, apparatus and system for facilitating data transport between WAN, SAN and LAN and for enabling enterprise computing into networks
US6633565B1 (en) * 1999-06-29 2003-10-14 3Com Corporation Apparatus for and method of flow switching in a data communications network
US20030204587A1 (en) * 2002-04-29 2003-10-30 Harris Corporation Tracking traffic in a mobile Ad Hoc network
US6707819B1 (en) * 1998-12-18 2004-03-16 At&T Corp. Method and apparatus for the encapsulation of control information in a real-time data stream
US6788681B1 (en) * 1999-03-16 2004-09-07 Nortel Networks Limited Virtual private networks and methods for their operation
US20040184408A1 (en) * 2003-03-22 2004-09-23 Sbc Properties, L.P. Ethernet architecture with data packet encapsulation
US6807172B1 (en) * 1999-12-21 2004-10-19 Cisco Technology, Inc. Method and apparatus for learning and switching frames in a distributed network switch
US20050174941A1 (en) * 2004-02-09 2005-08-11 Shanley Timothy M. Methods and apparatus for controlling the flow of multiple signal sources over a single full duplex ethernet link
US20050243818A1 (en) * 2004-04-30 2005-11-03 Infineon Technologies Ag Router IP port for an IP router
US20050276279A1 (en) * 2004-06-10 2005-12-15 Alcatel Network unit for exchanging protocol data units through tunnels
US20060146832A1 (en) * 2005-01-05 2006-07-06 CISCO TECHNOLOGY, INC. A Corporation of the state of California Method and system for transporting data using pseudowire circuits over a bridged network
US20060168274A1 (en) * 2004-11-08 2006-07-27 Eliezer Aloni Method and system for high availability when utilizing a multi-stream tunneled marker-based protocol data unit aligned protocol
US20060187948A1 (en) * 2005-02-18 2006-08-24 Broadcom Corporation Layer two and layer three virtual private network support in a network device
US20070071014A1 (en) * 2003-10-14 2007-03-29 Raptor Networks Technology, Inc. Switching System with Distributed Switching Fabric
US20070297411A1 (en) * 2006-06-23 2007-12-27 Huawei Technologies Co., Ltd. Method and Device for Generic Framing Procedure Encapsulation
US20080126607A1 (en) * 2006-09-20 2008-05-29 Integrated Device Technology, Inc. Method of handling flow control in daisy-chain protocols
US20090157860A1 (en) * 2007-12-14 2009-06-18 Raptor Networks Technology, Inc. Disaggregated network management
US20090161669A1 (en) * 2007-12-21 2009-06-25 Nortel Netowrks Limited Evolution of Ethernet Networks
US7593328B2 (en) * 2002-11-19 2009-09-22 Huawei Technologies Co., Ltd. Flow control method using feedback to control flow on the network, switch and channels in large communication networks
US7606229B1 (en) * 2002-11-08 2009-10-20 Cisco Technology, Inc. Generic bridge packet tunneling
US20100309924A1 (en) * 2007-12-20 2010-12-09 Neil Harrison Client/server adaptation scheme for communications traffic
US8477620B2 (en) * 2006-12-27 2013-07-02 Entry Point, Llc System and method to provide multiple private networks using PBB

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080114865A1 (en) * 2006-11-14 2008-05-15 Rothman Michael A Methods and apparatus to manage computing platforms
CN101796865B (en) * 2007-07-13 2013-05-01 北方电讯网络有限公司 Quality of service control in multiple hop wireless communication environments
GB2459838B (en) * 2008-05-01 2010-10-06 Gnodal Ltd An ethernet bridge and a method of data delivery across a network

Patent Citations (31)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5872783A (en) * 1996-07-24 1999-02-16 Cisco Systems, Inc. Arrangement for rendering forwarding decisions for packets transferred among network switches
US5999531A (en) * 1998-04-17 1999-12-07 Cabletron Systems, Inc. Method and system for identifying ports and forwarding packets in a multiport switch
US20010043609A1 (en) * 1998-04-28 2001-11-22 Alan Stanley John Chapman Efficient packet data transport mechanism and an interface therefor
US20020018475A1 (en) * 1998-07-22 2002-02-14 Yoram Ofek Switching with multiple time references
US6707819B1 (en) * 1998-12-18 2004-03-16 At&T Corp. Method and apparatus for the encapsulation of control information in a real-time data stream
US6788681B1 (en) * 1999-03-16 2004-09-07 Nortel Networks Limited Virtual private networks and methods for their operation
US6633565B1 (en) * 1999-06-29 2003-10-14 3Com Corporation Apparatus for and method of flow switching in a data communications network
US6807172B1 (en) * 1999-12-21 2004-10-19 Cisco Technology, Inc. Method and apparatus for learning and switching frames in a distributed network switch
US20030152182A1 (en) * 2000-08-22 2003-08-14 Pai B. Anand Optical exchange method, apparatus and system for facilitating data transport between WAN, SAN and LAN and for enabling enterprise computing into networks
US20020061018A1 (en) * 2000-10-03 2002-05-23 Altima Communications, Inc. Switch transferring data using data encapsulation and decapsulation
US20020083190A1 (en) * 2000-12-26 2002-06-27 Satoshi Kamiya Apparatus and method for GFP frame transfer
US20020101842A1 (en) * 2001-02-01 2002-08-01 Harrison Lee C. Interface system for wireless node and network node
US20020165978A1 (en) * 2001-05-07 2002-11-07 Terence Chui Multi-service optical infiniband router
US20030204587A1 (en) * 2002-04-29 2003-10-30 Harris Corporation Tracking traffic in a mobile Ad Hoc network
US7606229B1 (en) * 2002-11-08 2009-10-20 Cisco Technology, Inc. Generic bridge packet tunneling
US7593328B2 (en) * 2002-11-19 2009-09-22 Huawei Technologies Co., Ltd. Flow control method using feedback to control flow on the network, switch and channels in large communication networks
US20040184408A1 (en) * 2003-03-22 2004-09-23 Sbc Properties, L.P. Ethernet architecture with data packet encapsulation
US20070071014A1 (en) * 2003-10-14 2007-03-29 Raptor Networks Technology, Inc. Switching System with Distributed Switching Fabric
US20050174941A1 (en) * 2004-02-09 2005-08-11 Shanley Timothy M. Methods and apparatus for controlling the flow of multiple signal sources over a single full duplex ethernet link
US20050243818A1 (en) * 2004-04-30 2005-11-03 Infineon Technologies Ag Router IP port for an IP router
US7499450B2 (en) * 2004-04-30 2009-03-03 Infineon Technologies Ag Router IP port for an IP router
US20050276279A1 (en) * 2004-06-10 2005-12-15 Alcatel Network unit for exchanging protocol data units through tunnels
US20060168274A1 (en) * 2004-11-08 2006-07-27 Eliezer Aloni Method and system for high availability when utilizing a multi-stream tunneled marker-based protocol data unit aligned protocol
US20060146832A1 (en) * 2005-01-05 2006-07-06 CISCO TECHNOLOGY, INC. A Corporation of the state of California Method and system for transporting data using pseudowire circuits over a bridged network
US20060187948A1 (en) * 2005-02-18 2006-08-24 Broadcom Corporation Layer two and layer three virtual private network support in a network device
US20070297411A1 (en) * 2006-06-23 2007-12-27 Huawei Technologies Co., Ltd. Method and Device for Generic Framing Procedure Encapsulation
US20080126607A1 (en) * 2006-09-20 2008-05-29 Integrated Device Technology, Inc. Method of handling flow control in daisy-chain protocols
US8477620B2 (en) * 2006-12-27 2013-07-02 Entry Point, Llc System and method to provide multiple private networks using PBB
US20090157860A1 (en) * 2007-12-14 2009-06-18 Raptor Networks Technology, Inc. Disaggregated network management
US20100309924A1 (en) * 2007-12-20 2010-12-09 Neil Harrison Client/server adaptation scheme for communications traffic
US20090161669A1 (en) * 2007-12-21 2009-06-25 Nortel Netowrks Limited Evolution of Ethernet Networks

Non-Patent Citations (8)

* Cited by examiner, † Cited by third party
Title
Cisco, "Cisco Server Fabric Switch InfiniBand Fabric, White Paper", 11/06, 18 pages. *
Fujitsu, “The TCP/IP Protocol Suite”, Fujitsu, December 20, 2006, 74 pages. *
Hochmuth, "What's the biggest, fastest LAN switch", Nov 7, 2006, Computerworld, 7 pages. *
Mamakos, L. et al. (Editor), rfc2516 A Method for Transmitting PPP Over Ethernet (PPPoE), February 1999, 17 pages. *
Martini et al., “rfc4448 Encapsulation Methods for Transport of Ethernet over MPLS Networks”, Network Working Group, April 2006, 23 pages. *
Mellanox Technologies Inc, "Introduction to InfiniBand, White Paper", 2003, 20 pages. *
rfc4391, "Transmission of IP over InfiniBand (IPoIB)", Network Working Group, April 2006, 21 pages. *
Simpson, W (Editor), rfc1662 PPP in HDLC-like Framing, July 1994, Network Working Group, 25 pages. *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
EP3185491A1 (en) * 2014-02-05 2017-06-28 INTEL Corporation Reliable transport of ethernet packet data with wire-speed and packet data rate match
EP3103233A4 (en) * 2014-02-05 2017-08-09 Intel Corp Reliable transport of ethernet packet data with wire-speed and packet data rate match

Also Published As

Publication number Publication date Type
EP2291959B1 (en) 2017-04-05 grant
US9401876B2 (en) 2016-07-26 grant
CN102318283A (en) 2012-01-11 application
EP2291959A1 (en) 2011-03-09 application
WO2009133369A8 (en) 2010-12-16 application
US20160337146A1 (en) 2016-11-17 application
GB2459838B (en) 2010-10-06 grant
GB2459838A (en) 2009-11-11 application
WO2009133369A1 (en) 2009-11-05 application
GB0807937D0 (en) 2008-06-11 grant

Similar Documents

Publication Publication Date Title
US6954463B1 (en) Distributed packet processing architecture for network access servers
US7400590B1 (en) Service level to virtual lane mapping
US7869432B1 (en) Peer-to-peer link aggregation across a service provider network
US7616646B1 (en) Intraserver tag-switched distributed packet processing for network access servers
US7706255B1 (en) Communications apparatus with redundant switching or backpressure mechanism
US6571291B1 (en) Apparatus and method for validating and updating an IP checksum in a network switching system
US20070280223A1 (en) Hybrid data switching for efficient packet processing
US6587431B1 (en) Supertrunking for packet switching
US20060203721A1 (en) Apparatus and method for packet forwarding with quality of service and rate control
US6772267B2 (en) Multi-portal bridge for providing network connectivity
US20060140130A1 (en) Mirroring in a network device
US8351352B1 (en) Methods and apparatus for RBridge hop-by-hop compression and frame aggregation
US20020165978A1 (en) Multi-service optical infiniband router
US20120158942A1 (en) Methods and apparatus for forwarding-state transport in a distributed control plane
US20120179796A1 (en) Routing and service performance management in an application acceleration environment
US20040109473A1 (en) Interconnecting network processors with heterogeneous fabrics
US20020071438A1 (en) Network architecture and methods for transparent on-line cross-sessional encoding and transport of network communications data
US20060146832A1 (en) Method and system for transporting data using pseudowire circuits over a bridged network
US6510135B1 (en) Flow-level demultiplexing within routers
US20130051235A1 (en) End-to-end lossless ethernet in ethernet fabric
US5946308A (en) Method for establishing restricted broadcast groups in a switched network
US8369347B2 (en) Fiber channel over Ethernet and fiber channel switching based on Ethernet switch fabrics
US7599289B2 (en) Electronic communication control
US6697872B1 (en) Distributed packet processing using encapsulation and decapsulation chains
US20040151180A1 (en) Enhanced H-VPLS service architecture using control word

Legal Events

Date Code Title Description
AS Assignment

Owner name: GNODAL LIMITED, UNITED KINGDOM

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:BEECROFT, JON;HEWSON, DAVID CHARLES;FORD, ANTHONY MICHAEL;AND OTHERS;REEL/FRAME:025310/0681

Effective date: 20101105

AS Assignment

Owner name: CRAY UK LIMITED, UNITED KINGDOM

Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNOR:GNODAL LIMITED;REEL/FRAME:031681/0309

Effective date: 20131021